Recognizing Context for Privacy Preserving of First Person Vision Image Sequences

Abstract

The constant increasing evolution of life-logging wearable devices, as well as the fast grow of their market, has introduced relevant changes in the acquisition, storage and automatic understanding of images and videos. Along with the novel users’ opportunities, this technology is introducing a large amount of privacy-related concerns, mainly regarding the unaware or unwilling contexts subject that could get recorded by a life-logging device. In this work, we devise an approach to help life-logging wearable devices enforcing restrictions for context-related users’ privacy preservation. The proposed approach joins different technological innovations, from computer vision techniques to bluetooth beacon technology and network security.